The "We-Can-Make- A-Peer-Group" Valuation Fallacy (Fool on the Hill) January 13, 2000

An Investment Opinion

The "We-Can-Make- A-Peer-Group" Valuation Fallacy

By Bill Barker (TMF Max)
January 13, 2000

The results of baseball's Hall of Fame voting were announced on Tuesday, with Tony Perez and Carlton Fisk being the recipients of the honor of induction this year. That announcement reminded me to go back to my bookshelf and grab one of my best resources on stock market valuations, Bill James' The Politics of Glory: How Baseball's Hall of Fame Really Works.

The reason that I select this book (actually any book by Bill James would suffice) is that the author is so meticulous about using statistics and numbers in a logical and orderly way, and is so good at dissecting the fake arguments that are sometimes used by proponents of one cause or another. One of the arguments that James finds particularly misleading is what he calls, "The We-Can-Make-A-Group Argument." In relation to baseball, this is used by selecting two or more statistical categories in which an individual has statistically excelled, surrounding a player in a group of better and more well-rounded players, and ignoring all of the other data that give a truer picture a player's real accomplishments and worth.

The argument typically goes something like this: "Name the only man who achieved a career batting average above .320 with more than 5,000 at bats who has not been elected to the Hall of Fame." (Babe Herman is the answer.) The purpose of creating such an argument is that it is supposed to lead, inexorably, to the conclusion that if there are 30 players or so with names like Ruth, Williams, and Cobb who have this characteristic, and all but Herman are in the Hall of Fame, then, well, obviously Herman should go in.

The Wall Street research analyst variant of this strategy that you'll see far too often is the so-called "peer group" valuation. One of the sure signs that an analyst doesn't deeply believe what he's saying about the attractive current price tag on a company is the creation of a group, the components of which bear little, if any, relationship to the company being analyzed. I wrote about this phenomenon a couple of months ago in The Unbearable Silliness of Price Targets, and I've come across another pretty good example this week -- a December 13 Morgan Stanley research report on United Parcel Service (NYSE: UPS).

The report is an initiation of coverage on UPS, and puts a rating of "Outperform" and a "price target" of $80 on the company. The centerpiece of the valuation analysis is as follows: "To determine the appropriate valuation [of UPS] we screened other "Best in Breed" companies, narrowing our focus on a select group of blue chip, multinational companies with strong brand names and outstanding financial returns. This group includes General Electric, Procter and Gamble, The Home Depot, and Wal-Mart. These stocks currently trade at nearly 39 times 2000 estimates and about 32 times 2001 estimates. Our current target assumes that UPS can trade to a 10% discount to this peer group. That discount reflects the need, in our view, for the company's shares to season[.]"

In a long list of competitors in the category of Analyst's Most Egregious Act of Putting the Rabbit in the Hat, this has to rank close to the top. Morgan Stanley's analysis can be summed up as follows: "In comparison to companies in completely different businesses with high P/Es, the P/E for UPS is somewhat lower. Therefore it's attractively priced, and should sell for more."

It isn't hard to see why Morgan Stanley decided to discard any pretense of rigorous analysis in this particular instance. Since Morgan was the lead underwriter of the IPO, there was an understood obligation that Morgan would initiate its coverage with a buy rating and a price target above the level at which the company was trading -- basically no matter what the then price of the company was.

The problem for the analyst obligated to produce an attractive price target in this case was that the market had already pushed UPS shares up awfully high following its IPO. At the December 13 price of $66 a share, the level was well above the IPO price, a price which itself had been raised more than once prior to the big IPO day. In its October 20, 1999 S-1/A registration, UPS indicated that it expected to receive between $36 and $42 a share for the 109.4 million shares being sold at its initial public offering. That range was subsequently raised, and by its November 5 S-1/A, UPS indicated that the expected range was $47 to $49 a share. Ultimately the company came public at $50 a share.

It's fun to speculate on what the implications of an $80 per share "appropriate valuation" would be if we were to take it at face value. It would appear that the claim would have to be not only that the market doesn't quite know what it's doing by pricing UPS at a mere $66 a share, but that anybody who would think that a reasonable price could even be $50, or, Mon Dieu, $36, really is kind of a bumbling idiot. That is, if the "appropriate valuation" technique prepared by Morgan Stanley's analyst were in any sense really appropriate, should the Morgan Stanley bankers who initially priced the IPO at $42 or less all be fired for incompetence? Check out Bill Mann's comments on this phenomenon from yesterday's column.

Ultimately, the big problem here with Morgan's analysis is why was the "peer group" composed of the companies that were chosen? The companies mentioned don't, to me, bear any close resemblance to UPS in terms of business models, or anything else that I'm aware of. They appear to have been selected simply on the basis of their high P/Es. Put another way, would you think that the peers for evaluating a transport company would be retailers, a household products company, and whatever you choose to classify GE as?

Why are these companies more like UPS (and therefore better predictors of the future or appropriate P/E of UPS) than, say, Merrill Lynch, Boeing, Kodak, and Federal Express? Those stocks trade at about 14 times 2000 estimates, but they are all "blue chip, multinational companies with strong brand names and outstanding financial returns," aren't they? (Well, maybe Kodak doesn't really have outstanding returns.) What about Microsoft, Cisco, The Gap, and America Online? Those also are "blue chip, multinational companies with strong brand names and outstanding financial returns," and they've got average P/Es of closer to 100. Is that how UPS should trade?

Morgan's analysis doesn't give any guideline as to what makes one "blue-chip" company more appropriate than another to determine the "appropriate valuation" for UPS -- but then actually arriving at an appropriate valuation for UPS isn't really the analyst's goal in the first place, despite the statement to the contrary. The goal was simply to give some sort of justification to a higher price target.

The deeper problem with these types of analyses -- the preselection of outstanding companies as "peers" for other companies -- is that every single company in the market thus becomes undervalued through this method, because there are always going to be vaguely comparable companies that carry higher price tags than others. Even the companies with the highest pice/earnings ratios can still be compared to other companies that have higher price/sales or price/earnings/growth ratios. It never ends. Ultimately, through the use of inappropriate peer groups, no company at all is undervalued, or even fairly valued. Ever.

In this sense, tossing UPS into a group with Home Depot isn't really the problem. They're both good companies and people aren't going to lose their shirts buying UPS at $66 a share. (Indeed, I'll be arguing the bull side in a Dueling Fools piece on UPS next week.) The problem is that any company, no matter how awful, can be tossed into a nonsensical peer group. in a group with Yahoo!? I've seen that. If you're willing to strain the definitions of what defines a peer so that size, profitability or even business sector are not determinants or even guides to the components of the group, you've chosen the conclusion first, and then selected companies to support your thesis second.

That, of course, will never help anyone find what the "appropriate valuation" of a company is.